Bug#712237: cups-server-common: The cost factor for pstops

[Till Kamppeter <till.kamppeter@gmail.com>]

The old cost factor (65) I had introduced to overcome an ugliness in the
PDF-based printing workflow.

If the cost factor is 66 and an app sends a PostScript input file and
the printer is PostScript (or some old-fashioned driver insisting on
PostScript input is used), I got

PostScript -> pstopdf -> PDF -> pdftopdf -> PDF -> pdftops -> PostScript

a detour of PostScript being converted to PDF and back to PostScript
which happened sometimes because some apps (like emacs) did not switch
over to the PDF-based printing workflow sending out print data in PDF.

This principally works but causes delays (and even failures) due to
unneeded CPU- and meomory-intensive re-rendering and also sometimes font
problems as in bug #712015.

So I created a bypass for this situation by using the 65 cost factor
resulting in

PostScript -> pstops -> PostScript

This worked well until CUPS 1.6.x came up where this workaround caused
"make test" to fail and so the build of the Debian and Ubuntu packages.
Therefore we reverted to 66.

The best would be getting all apps sending PDF (and cups-pdf accepting
PDF), but as this will probably take years (and there are also apps in
use which are not maintained any more), we should better fix the "make
test" so that the tests pass also with the 65 cost factor or modifying
the filter workflow generally does not break the tests.

   Till


On 06/14/2013 01:12 PM, Brian Potkin wrote:
> Package: cups-server-common
> Version: 1.6.2-8
> Severity: normal
> 
> 
> In /usr/share/cups/mime/mime.convs there is
> 
>    application/postscript          application/vnd.cups-postscript 66      pstops
> 
> In cups 1.5.x we have
> 
>    application/postscript          application/vnd.cups-postscript 65      pstops
> 
> Unless it is intended to alter the work flow for all PostScript documents
> I'd consider this is a bug. There is nothing obvious in the changelog
> to account for the change. (But it not unknown for the significance of
> something to pass me by :) ).
> 
> Regards,
> 
> Brian.
> 
> 
> 
> 
> -- System Information:
> Debian Release: 7.0
>   APT prefers stable
>   APT policy: (500, 'stable')
> Architecture: i386 (i686)
> 
> Kernel: Linux 3.2.0-4-686-pae (SMP w/1 CPU core)
> Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)
> Shell: /bin/sh linked to /bin/dash
>